Engineering case studies
Independent systems and professional capabilities presented through the problems, constraints, architectural decisions, trade-offs, validation, and outcomes that shaped the work.
Open source · AGPLv3
A desktop app that finds cloud services you are paying for but not using, like forgotten disks, idle IP addresses, and old projects. Your keys stay on your computer.
Working today Checks AWS, Google Cloud, Azure, Vercel, Supabase, and OpenAI. Every dollar estimate shows its math, and nothing is deleted without your confirmation.
FEATURE_NEW_AUTHfalse → true
JWT_ALGORITHMHS256 → RS256
Open source · Marketplace
Give it a working config, a broken config, and a command that shows the failure. It finds the smallest set of changes that still breaks things, including failures that only appear when two settings interact.
Published today v0.4.1 on PyPI and in the GitHub Marketplace. Handles .env, JSON, YAML, and TOML, and runs candidates locally, in Docker, in Kubernetes, or in CI.
pipx install envcause
Live on Chrome Web Store
A local-first Chrome extension that shows which of your other extensions can reach ChatGPT, Claude, Gemini, and other AI sites, and cites the exact permission behind every finding.
Live today Free on the Chrome Web Store. Seven AI services, four exposure levels, change tracking between scans, one-click disable, and Markdown/JSON reports. Two permissions, zero network code.
C#/.NET · Angular · Azure
A job-search tracker built end to end to show a current .NET stack: an Angular board, a JWT-protected ASP.NET Core API, SQL Server, and a scheduled Azure Function.
Working today Multi-user app with per-user data isolation, search and pagination, API integration tests, one-command Docker setup, and Azure deploy pipeline. The follow-up function logs candidates; it does not send email yet.
Patterns across 12 years of delivery
Complex Angular applications, Node.js services, enterprise PHP platforms, integrations, reporting, and production workflows owned from requirements through ongoing support.
Learning undocumented systems, resolving production backlogs, modernizing code safely, improving database performance, and preserving business continuity.
AWS and Google Cloud deployments, Docker, CI/CD, logging, monitoring, and Python or PowerShell tools that reduce repetitive investigation and release work.